options(rgl.useNULL = TRUE)

pkg.req <- c("shiny", "shinyRGL", "rgl")

for (pkg in pkg.req) require(pkg, character.only = TRUE, quietly = TRUE, warn.conflicts = FALSE)

shinyServer(function(input, output) {

rootreps = 5; reps = rootreps ^ 2;

output$perspPlot <- renderWebGL({

# Creating Data for plotting using given replicates

x1 <- rep(c(-1, -1, 1, 1), each = reps)

x2 <- rep(c(-1, 1, -1, 1), each = reps)

x1x2 <- x1 * x2

# Design Matrix computer simulation

ones <- rep(1, 4 * reps)

design <- cbind(ones, x1, x2, x1x2)

# Sets fixed seed for noise simulation

set.seed(10)

err <- rnorm(4 * reps, 0, 1)

errors <- err * input$sigma

# Model parameters

coefs <- matrix(c(input$mu, input$alpha, input$beta, input$alpha.beta), ncol = 1)

# Simulate data

y <- design %*% coefs + errors

# Ordering data for plotting

yarr <- array(0,c(rootreps, rootreps, 4))

for (i in 1:4) {

yarr[,,i] <- matrix(y[((i - 1) * reps + 1):(i * reps)], rootreps, rootreps)

}

ymat <- rbind(cbind(yarr[, , 1], yarr[, , 2]), cbind(yarr[, , 3], yarr[, , 4]))

# Prespective Plot

z1 <- seq(0, 1, 1 / (2 * rootreps - 1))

persp3d(z1, z1, ymat, expand = 0.5, col = "green",

xlab = "A", ylab = "B", zlab = "y",

ticktype = "detailed", nticks = 2, zlim = c(3,7))

})

})